gm灰色模型matlab源程序代码

function yc0=GM1(x0,N) n=length(x0); for i=1:n x1(i)=sum(x0(1:i)); end for k=2:n % 紧邻均生成 z % z(k)=0.5*x1(k)+0.5*x1(k-1); z(k)=((x1(k)-x1(k-1))/(log2(x1(k))-log2(x1(k-1))))*log2(exp(1)); end % for i=1:n-1 % b(i,1)=-z(i+1); % y(i)=x0(i+1); % end % b(:,2)=1; b(:,1)=-z(2:n)'; b(:,2)=1; y=x0(2:n); y=y'; % 转置为列向量 au=b\y; % 作矩阵除法，计算a u yc1(1)=x0(1); d=(1-exp(-n*au(1)))/(1-exp(-au(1))); c=(sum(x1)+(d-n)*au(2)/au(1))/d; for k=1:n+N-1 % c=x0(1)-au(2)/au(1); yc1(k+1)=(c-au(2)/au(1))*exp( -au(1)*k)+au(2)/au(1); end yc0(1)=x0(1); % for k=1:n+N-1 % yc0(k+1)=yc1(k+1)-yc1(k); % end yc0(2:n+N)=yc1(2:n+N)-yc1(1:n+N-1); % disp(uint16(yc0(2:1:n+1))); % for k=1:n % e0(k)=x0(k)-yc0(k); % end e0=x0(1:n)-yc0(1:n); max1=max(abs(e0)); r=1; for k=2:n r=r+0.5*max1/(abs(e0(k))+0.5*max1); end r=r/n; % r 表示关联度 pe=mean(e0); % px0=sum(x0)/n; % z=x0-px0; % S1=sqrt(sum(z.^2)/n); % z=e0-pe; % S2=sqrt(sum(z.^2)/n); % C=S2/S1 c= std(e0)./std(x0) p=length(find(abs(e0-pe)<0.6745*std(x0)))/n; if p>0.95&c<0.35 disp (' 预测精度好'); else if p>0.8&c<0.5 disp (' 预测合格'); else if p>0.7&c<0.65 disp (' 预测勉强合格'); else disp (' 预测不合格'); end end end